Search Results for author:
Found 13 papers, 9 papers with code
Learning Important Features Through Propagating Activation Differences
Here we present DeepLIFT (Deep Learning Important FeaTures), a method for decomposing the output prediction of a neural network on a specific input by backpropagating the contributions of all neurons in the network to every feature of the input.
Not Just a Black Box: Learning Important Features Through Propagating Activation Differences
Note: This paper describes an older version of DeepLIFT.
Computationally Efficient Measures of Internal Neuron Importance
We compare Neuron Integrated Gradients to DeepLIFT, a pre-existing computationally efficient approach that is applicable to calculating internal neuron importance.
WILDS: A Benchmark of in-the-Wild Distribution Shifts
Distribution shifts -- where the training distribution differs from the test distribution -- can substantially degrade the accuracy of machine learning (ML) systems deployed in the wild.
Technical Note on Transcription Factor Motif Discovery from Importance Scores (TF-MoDISco) version 0.5.6.5
TF-MoDISco (Transcription Factor Motif Discovery from Importance Scores) is an algorithm for identifying motifs from basepair-level importance scores computed on genomic sequence data.
A General Framework for Abstention Under Label Shift
In this work, we present a general framework for abstention that can be applied to optimize any metric of interest, that is adaptable to label shift at test time, and that works out-of-the-box with any classifier that can be calibrated.
Maximum Likelihood with Bias-Corrected Calibration is Hard-To-Beat at Label Shift Adaptation
Label shift refers to the phenomenon where the prior class probability p(y) changes between the training and test distributions, while the conditional probability p(x|y) stays fixed.
Expectation Maximization with Bias-Corrected Calibration is Hard-To-Beat at Label Shift Adaptation
A limiting assumption of this algorithm is that p(y|x) is calibrated, which is not true of modern neural networks.
Network modelling of topological domains using Hi-C data
However, one of the drawbacks of community detection is that most methods take exchangeability of the nodes in the network for granted; whereas the nodes in this case, i. e. the positions on the chromosomes, are not exchangeable.
Applications Genomics
Unsupervised Learning from Noisy Networks with Applications to Hi-C Data
In this paper, we propose an optimization framework to mine useful structures from noisy networks in an unsupervised manner.
The importance of transparency and reproducibility in artificial intelligence research
In their study, McKinney et al. showed the high potential of artificial intelligence for breast cancer screening.
Applications
Fourier-transform-based attribution priors improve the interpretability and stability of deep learning models for genomics
To address these shortcomings, we propose a novel attribution prior, where the Fourier transform of input-level attribution scores are computed at training-time, and high-frequency components of the Fourier spectrum are penalized.
Adapting to Label Shift with Bias-Corrected Calibration
Label shift refers to the phenomenon where the marginal probability p(y) of observing a particular class changes between the training and test distributions, while the conditional probability p(x|y) stays fixed.
